Lennart Heim is an information scientist and researcher in AI governance at the RAND Corporation and a leading scholar on AI export controls. We asked him about the Biden administration’s “diffusion framework,” which aims to regulate the global diffusion of advanced AI chips and models. We get into all the specifics as well as the broader geopolitical implications of the framework—and whether or not the Trump administration will maintain this policy.
Detailed Summary
In January 2024, during the final days of the Biden administration, the Commerce Department's Bureau of Industry and Security released a significant new framework called the "diffusion rule" to regulate global AI infrastructure. This rule built upon previous export controls from both the Trump and Biden administrations that had targeted Chinese access to advanced semiconductors and AI chips.
The diffusion rule established a three-tier system for global AI infrastructure access. The US and its closest allies received the most freedom, while countries like China, Russia, and North Korea faced the strictest restrictions. Most nations fell into a middle tier with regulated access based on quantity thresholds and licensing requirements.
For these middle-tier countries, smaller deployments (under 1,700 NVIDIA H100-equivalent chips) faced minimal restrictions, while larger deployments required increasingly stringent oversight. The largest deployments needed special "national verified end user" status and had to meet strict requirements around cybersecurity, model security, and business relationships.
The discussion highlighted key tensions in this approach. While currently NVIDIA's chips maintain such a technological lead that most companies accept these restrictions to access them, this advantage might not last. If Chinese alternatives become competitive, the regulatory burden could push countries away from US technology. This risk becomes particularly acute if AI development takes longer than expected - the rules appear optimized for a world where transformative AI emerges within 3-5 years.
Looking forward, while the Trump administration could modify these rules, bipartisan support for restricting Chinese AI development suggests some version will likely continue. The key question remains whether these controls will successfully preserve US technological advantage or ultimately undermine it by accelerating development of alternative ecosystems.
Share this post